CN1846210A - Method and apparatus for storing and retrieving data using ontologies - Google Patents
Method and apparatus for storing and retrieving data using ontologies Download PDFInfo
- Publication number
- CN1846210A CN1846210A CNA2004800248868A CN200480024886A CN1846210A CN 1846210 A CN1846210 A CN 1846210A CN A2004800248868 A CNA2004800248868 A CN A2004800248868A CN 200480024886 A CN200480024886 A CN 200480024886A CN 1846210 A CN1846210 A CN 1846210A
- Authority
- CN
- China
- Prior art keywords
- node
- term
- document
- verb
- noun
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3344—Query execution using natural language analysis
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
A method of storing service description documents in a computerised storage system in which each document is associated with at least one verb ontological node (204) and at least one noun ontological node (254), each verb ontological node having one or more links to other verb ontological nodes and each noun ontological node having one or more links to other noun ontological nodes whereby the verb nodes form a verb space (200) and the noun nodes form a noun space (250) and a method of retrieving service description documents from a plurality of service description documents stored in this way comprising the steps of: controlling a user interface to request from a user at least one verb request term (405) and at least one noun request term (410), associating the or each verb request term (405) with a corresponding verb node (204) and the or each noun request term (410) with a corresponding noun node (254), comparing the or each corresponding verb node (204) with the or each verb node (212, 214) associated with each of the stored service description documents, comparing the or each corresponding noun node (254) with the or each noun node (262, 266) associated with each of the stored service description documents, and selecting for retrieval zero or more of the stored service description documents on the basis of the comparison steps and controlling the user interface to inform the user of the selected documents to enable the user to retrieve one or more of the selected documents.
Description
Technical field
The present invention relates to such method and apparatus, that is, it is used for storage and retrieve data, more specifically is used to store and retrieve the data that the description of the user's that particularly offers described method or device service is represented.
Background technology
At present, have the Different Strategies of limited quantity, it is used to store lot of data, makes the user to search in the data of being stored then, to retrieve the data of being paid close attention to.Can be with these policy class for falling into certain position between the two poles of the earth (extreme), the utmost point in these the two poles of the earth is represented that by WWW another is extremely represented by the tree construction catalogue.
In the former, there is not the structure of the mode that adopts the storage data; In order in the data of being stored web page or leaf or " document " of a large amount of texts (and particularly include usually), to search for, use search engine usually based on keyword.
A kind of very simple keyword search engine can only be to search to drag for (trawl) (perhaps " climb and seek (crawl) ") with the just searched keyword of searching in the document of storage, and returns all this documents that comprise described keyword.Yet, handle in order when having the document of a large amount of storages, to accelerate, a kind of more complicated search engine can generate index in advance, and this index comes the document of all storages is set up index according to certain keyword frequency that (the document is given mark through pre-service and at each " keyword " of occurring in the document) occurs in document.Thereby this search may further comprise the steps: index is searched for, and return those search of carrying out at the frequency of occurrences fully high document, the perhaps identifier of these documents of keyword.
This tactful shortcoming is: keyword may have more than one implication, and therefore search may be returned (corresponding with undesirable implication of keyword) irrelevant document.Similarly and since document used with searching request in the different term of employed term, so this search may be missed relevant documentation.
Yet when with logical order arrangement mode (for example with the sort tree structure catalogue) storage document, the user can search for this catalogue simply, stores till the leaf node that is considered to relative document up to arrival.This policies using the problems referred to above relevant with keyword search.
Yet, except each document being placed on the required effort in its tram in catalogue, this tactful shortcoming also is: if the searchers searches for along the route different with the contemplated route of the people who at first stores the document in directory tree, then this searchers may can not find relevant document.
In recent years, carried out a large amount of research for the use of body (ontology).Usually use body to help integrate allos and leave over database.Body is realized this purpose by the physical meaning that accurately is defined in the term that uses by different way in the different databases.For example, a database may be quoted " model ID ", and another database may be quoted the classification identical with " product ".Can use Ontology Mapping (mapping) that " model ID " in first body is mapped to " product " in second body then, by translating second body etc. from first body, make that this searching request can also be from another database retrieval related data thus by the searching request of first body statement.
In order to improve accuracy, also the possibility of utilizing " body " has been carried out a large amount of research in recent years according to the search of carrying out based on keyword method.As mentioned above, body be to usually in limited knowledge domain a kind of form of the implication of employed various terms represent.An example of known body is speech net (WordNet).This body is attempted to represent all English languages in formal mode.Each word has one or more possibility implication that is associated with it, and each implication then is linked to other implications of other words with different ways (for example, hyponym, superordinate term or the like).A kind of mode of utilizing body to improve the search accuracy is: the keyword in any or all two in searching request and the destination document is replaced with implication according to the keyword of body.Can use this main body structure then, include the document of the implication identical but also search for its hyponym etc., improve the dynamics of search with the implication of this request by not only searching for.(hyponym is a kind of particular type or the specialized type of things, and for example, throwing blade and fish knife are the hyponyms of cutter, and cutter is the superordinate term of fish knife etc.).The paper of discussing this method has: " OntoSeek:Content-Based Access to the Web " by Nicola Guarino, Caudio Masolo and Guido Vetere; Published in IEEE INTELLIGENTSYTEMS publication MAY/JUNE 1999 edition pages, 70 to 80.
A kind of commercially important application that is used to store document and makes the user to search for the document of being stored is to make that serving the provider can find the client, and vice versa.In this application, serve the provider in catalogue or similarly store in the facility their service advertisement of publication, and user or potential customers can search directories, serve the provider to attempt to find one or more that desired service can be provided.In this application, respectively serve the provider natural language description to its service that can provide or various services is provided usually.Utilize computerized system, the user then can carry out keyword search to these natural language descriptions, to attempt finding the suitable provider that serves.
Summary of the invention
According to a first aspect of the invention, provide a kind of method that the document of a plurality of storages is searched for, this method may further comprise the steps: store described a plurality of document; The expression of storage body, described body comprise a plurality of nodes and be divided into the subspace of at least two uniquenesses of being mutually related; At in described a plurality of documents each, at least one of the node of storage and first unique subspace of body is related, with related with at least one of the node of second unique subspace of body; The control user interface so that the user can use free text input to import nearly at least two search terms, and can be searched for each term and is associated with corresponding unique subspace of body; The search term of each input is only compared with the node of corresponding subspace, to attempt determining that one or more may mate or the part coupling; And may mate or the node of part coupling based on each, and related between the node of the document of described storage and body selected the document of one or more described storage, to present to the user.
Should be understood that this method is intended to carry out on the computer equipment of some kind.For example, when submitting the application in the existing typical calculation machine equipment, described storing step can be carried out on the digital storage media such as hard disk drive, and the step of described control user interface and described comparison step and described selection step can be by the suitably execution such as computer processing system of programming.In fact, said method has solved by the Computer Storage document and has not used human keeper to store the problem that also search file produced.
Computing machine is being very clumsy aspect the linguistic context (context) of determining free text input, and the result produces pseudo-result through regular meeting when allowing free text to import.By body being provided a plurality of unique subspaces (wherein can not mate the node that produces in another unique subspace) as coupling or part for the search of the node in the sub spaces, and make the user in unique subspace, to search for individually, can alleviate the problem of clumsy linguistic context to a certain extent, still provide the facility that to import free text key word type search term simultaneously to the user.
Thus, the document request at human librarian of employing " I ' mihterested in finding out documents aboutbuilding restaurants and people providing restaurant buildingservices " statement is blured, and human librarian perhaps can be found out relevant document with a wet finger.But for computer system, same request perhaps (possibility) follows the minority relevant documentation to select a large amount of uncorrelated documents together.A problem of this request is: term " building " more is commonly used for noun, rather than the verb of being envisioned in this example.In one embodiment of the invention, existence is used for two nodes of " building ", one in the verb subspace of body, one in the noun subspace, etc.
According to a second aspect of the invention, provide a kind of in the computerize storage system stores service method of document is described, wherein each document is associated with at least one this body node of verb and at least one this body node of noun, each this body node of verb has with one or more of other these body nodes of verb and links, and each this body node of noun has with one or more of other these body nodes of noun and links, thereby these verb nodes form the verb space, and these noun nodes form the noun space.
Term " service description document " is used for representing to adopt the data of any form, after it passes through suitable treatments where necessary, comprises that the mankind of the specific service that the side to " having " this service description document provides or machine readable describe.Attention: this service can be the service that provides to another machine of a machine (for example is used for treating by automatic search and request the substep of the complex task carried out by miscellaneous equipment, makes " agency (agent) " (promptly with certain autonomous degree operation computer program) can carry out complex task).
Preferably, this method is further comprising the steps of: one in each service description document and a plurality of possible different relationship types is associated, and described possible different relationship types are represented the relation between each pairing of described at least one verb node and described at least one noun node.Preferably, described possible relationship type is: " ... input " (input-of), wherein, the service of in service description document, describing with the input of the corresponding object of noun node that is associated as this service; " ... output " (output-of), wherein, the service of in service description document, describing with the output of the corresponding object of noun node that is associated as this service; Perhaps " being associated with ... " (related-to), as if it be at not specifying other relations or above-mentioned option default relationship of inappropriate situation all.
Preferably, the quantity of service description document verb node associated therewith is 1, and the quantity of noun node is 1 or 2.
According to a third aspect of the invention we, provide a kind of from a plurality of service description document of storage according to a second aspect of the invention retrieval service the method for document is described, this method may further comprise the steps:
The control user interface, asking at least one verb request term and at least one noun request term from the user,
Each verb request term is associated with corresponding verb node, and each noun request term is associated with corresponding noun node,
The verb node that each is corresponding compares with each verb node that is associated with the service description document of each storage,
The noun node that each is corresponding compares with each noun node that is associated with the service description document of each storage,
Select retrieving the service description document of zero or more a plurality of storages based on described comparison step, and the document of control user interface to notify the user to choose, make the user can retrieve the document that one or more is chosen.
Preferably, also this user interface is controlled, obtaining to specify the noun request term imported and the information of the relationship type between the verb request term from the user, and each incidence relation type of the service description document of this information and each storage compared.
The present invention provides significant advantage with respect to the current known method of carrying out document searching and retrieval.Particularly, use verb and noun or noun, verb and noun as the directly perceived simple searching request of searching request formation, to generate and to provide a large amount of information at the user.In addition, for the purpose of the service of searching, this is the specially suitable form that searching request will adopt.In addition, by (promptly with each document and different body plenum, verb space and noun space) at least two nodes be associated, and each document is associated with relation between each noun node and the verb node, can form the description very accurately to service, this description remains very directly perceived and easy prediction for unprofessional user.
More at large, the present invention also is associated the document of being stored with two (or more a plurality of) unique identifier terms, and unique body plenum of each unique identifier term and its oneself is associated.Can carry out file retrieval efficiently by in described uniqueness (with specific) body plenum, carrying out two search that (or more a plurality of) are independent thus.By making these body plenum be exclusively used in the type of identifier, to compare with using general body plenum, they will comprise node still less, and therefore search is more likely accurate.But, can under the situation that does not lose accuracy, increase the scope of the different unique way of description/sign document by having body plenum (thereby search also is like this) more than one.A part that link type between the various identifier terms is included as sign/search strategy has also increased possible accuracy and " resolution ".Attention: it is useful especially using specific body plenum (it may be the limited body plenum of term, perhaps abbreviates limited body as), and this is because this limited body provides pin-point accuracy level and dirigibility level for index and search.
According to a forth aspect of the invention, provide a kind of system, it is used for making can store a plurality of documents, and can search for and carry out the selectivity retrieval these a plurality of documents subsequently, and described system comprises: data store is used to store described a plurality of document; A plurality of data item, the node in the expression body, described node are classified at least two son groups of intrinsic two the unique subspaces of expression; And at each document in the document and a son group node and the node of the document in organizing with another son between at least two related; Described system also comprises user interface layout (arrangement), this user interface is arranged and is used to make the user can search for one or more document among the document of described a plurality of storages, described user interface is configured such that the user can utilize free text input to import one or more search term, and each can be searched for term and be associated with an intrinsic unique subspace; And described system also comprises the processor layout, this processor is arranged and can be operated that (described data item is represented with lower node so that each search term and data item are compared, promptly, what described node was classified into body will search in the identical subspace, subspace that term is associated with the user), with may mating or the part coupling of one or more node in identification and the subspace, and, select and return to document as the result who searches for according to and the node that is complementary of the search term of user input and the node that is associated with the document of described a plurality of storages.
Preferably, for except that the document of the node direct correlation of selecting and be matched with the search term, also select the document that is associated with the close association node of node that matches the search term, relation (comprising born of the same parents' relation, subrelation and father's relation) between the node in unique body subspace has been used in the selection of document.
According to a fifth aspect of the invention, as described in claim 11, provide a kind of data store that is used for the system of third aspect present invention.
Other preferred features of the present invention in appended dependent claims, have been set forth.
Description of drawings
For the present invention may be better understood, now with reference to accompanying drawing, only the form with example is described embodiments of the invention, in the accompanying drawings:
Fig. 1 is the The general frame of the embodiment of the invention;
Fig. 2 is the illustrated embodiment diagrammatic sketch of the graphic user interface that adopts on the user terminal of the embodiment of Fig. 1;
Fig. 3 is the block diagram of ontology server that illustrates in greater detail the embodiment of Fig. 1;
Fig. 4 is the schematically illustrating an of part that is stored in the body on the ontology server of Fig. 3, and it comprises two links, and each link links two nodes of body, and is associated with service describing in the storage data store in the embodiment in figure 1;
Fig. 5 will " refurbish (trimming) " moves schematically illustrating of four kinds of different modes that node and " house (house) " Object node link according to the method for expressing that uses in an embodiment of the present invention;
Fig. 6 is and similar the schematically illustrating of the Fig. 4 that is added with this figure, and how the node that is used for being illustrated in search procedure searching request and body mates; And
Fig. 7 is selecting and the process flow diagram of the method for match retrieval service describing based on input request according to the embodiment of the invention.
Embodiment
The overall illustration of Fig. 1 be used to store and device that retrieval service is described.This device comprises user terminal 5, ontology server 10 and data store 15.In operation, the specific graphic user interface (it being carried out more detailed description) that the user utilizes user terminal to use hereinafter with reference to Fig. 2, with controlled way to a certain degree to user terminal 5 input service describing searching request.Then, user terminal 5 sends to this request the ontology server 10 that is connected with user terminal by data communication network.Ontology server 10 (by the following mode of describing in further detail with reference to Fig. 3 to Fig. 7) is handled this request, and zero or the more a plurality of match service selecting to be stored in the data store 15 are described.Then selected service describing is sent to user terminal, this user terminal is shown to the user with these service describings as the result of user search queries.
Referring now to Fig. 2, the user interface of Shi Yonging comprises two text input frames 51,52 in the present embodiment, please the user import noun and verb (note: these are simple " free text " input frames, and the user can key in any keyword and unrestricted (for example must select) as required from allow lists of keywords therein) respectively therein.Be provided with three radio buttons 53,54,55 below text frame, they are denoted as " ... input ", " ... output " and " being associated with ... " respectively.These radio buttons have following attribute (this point is known in the art): the once necessary definite button of selecting in office, like this, if the user has chosen a radio button different with the current button of choosing, the then current button of choosing automatically becomes not selected.In the present embodiment, acquiescence has been chosen " being associated with ... " button.In addition, user interface in the present embodiment comprises some texts, should how to form searching request with guides user, the text is " at this input noun ... [noun text box 51]; at this input verb ... [verb text box 52]; select in the following button then, to indicate the relation between this noun and this verb.For example,, import in ' house ' and ' construction ' in order to search for the company that the house building service is provided, select then ' ... output ' radio button ".
In the present embodiment, user interface also be included in following exemplary text " following result has been returned in your search: ... " the result space 57 of below shows the service describing of choosing that this equipment generates to the user therein.
Referring now to Fig. 3, the ontology server 10 in the present embodiment schematically is depicted as and comprises I/O unit 105, system bus 110, processor arrangement 115 and system storage 120.As known in the prior art, this system bus interconnects other primary clusterings 105,115,120 of server, makes any one assembly to communicate with any other assembly.I/O unit 105 makes server to receive searching request from user terminal 5, and Search Results is returned to user terminal 5 under the control of processor 115, and from data store 15 reading of data.Storer 120 (also storing the program of the integrated operation that is used for control body server 10) also comprises and is used to store the body memory block 122 of body and the service chaining memory block 124 that is used to store a plurality of service chainings, below will be explained in more detail.
The reader that is skilled in technique will be clear and definite: Fig. 3 and be the very upper expression of server for the foregoing description of Fig. 3, wherein for clarity sake omitted the details of server computer wittingly.Except situation as herein described, server can be its structure and operate in and be known in the art any suitable traditional server computing machine.
Fig. 4 illustration be stored in the part of the body in the ontology server 10.Each frame (202 to 220 and 252 to 268) that comprises word is represented a node of this body, and each single line arrow is represented by the relation between the node of its connection, and each double-lined arrow is represented by the link between the node of its connection.The node that is connected to each other (202 to 220) at this figure left-hand side is verb (being also referred to as an action) node, and is noun (being also referred to as object) node at the node (252 to 268) that this figure right-hand side is connected to each other.Double-lined arrow links the noun node link to the verb node.
Illustrated in the keyword as a plurality of frames belows that are connected of the illustration of this body part, single arrow relation line (305) expression superclass relation/subclass relation, node wherein that arrow is pointed be appointed as with respect to arrow the subclass of node dorsad; For example, " Get " (204) and " Give " (206) are appointed as the subclass of " Transact " (202).Be used for the basis that a node is appointed as the subclass of another node be: also fall into by in the specified notion of Section Point if fall into by each specific example situation in the specified notion of first node, but all the interior specific example situations of notion that fall into Section Point all do not fall in the notion of first node yet, then first node can be appointed as the subclass of Section Point.
Double-head arrow relation line (310) expression similar relation promptly, will be appointed as similar each other by two nodes that this relation line connected; For example, " Acquire " (208) and " Buy " (214) are appointed as are had similar relation each other.Be used for two nodes are appointed as similar each other basis and be: each the interior specific example situation of notion that falls into a node also should fall in the notion of another intranodal, and vice versa.Attention: this is to be determined for the purpose of special body by the body deviser, and in the present embodiment, this purpose is to make the expection client and provide the company of desired service to interrelate.In the present embodiment, the deviser judges thus, for these purposes, Acquire and Buy are synonyms, even for other these speech of purpose can represent different notions (possible Buy be Acquire subclass (differing because people can obtain article disputablely is decided to be its payment, and people can not buy article and do not pay the bill for these article)).
Single arrow (but two-wire) link (315,320) expression " ... input "/" ... output " link.The link (315,320,325) (representing) by double-lined arrow always with the verb node link to the noun node.If the arrow of " ... input "/" ... output " link always points to the verb node from the noun node, then this link is that " ... the input " link (320) as input of noun node is adopted in the service that is associated of expression or service search request.On the contrary, if this arrow points to the noun node from the verb node, then this link is that the service that is associated of expression or service search request produce " ... the output " link (315) as output of noun node.The two-wire link of no arrow is peer link (325), service that its expression is associated or service search request are with the verb node that linked and noun node in unspecified mode (promptly, as the input, output or input and output the two, any other situation of any in perhaps in fact object can not being thought to input or output) be relative to each other.
Fig. 5 b to 5c illustration can be in the present embodiment with two node links, three kinds of possible modes together, by these three kinds of possible modes the service of each storage or service search request are associated with only comprising a noun node, verb node and one the linking of relation between the two.Fig. 5 a illustration can be how in alternative embodiment with these two node links to together, wherein each that is associated with the service or the service search request of storage link and can both comprise that input noun node also comprises and export the noun node.Fig. 5 a shows following situation thus: in alternative embodiment, noun node " House " be the input of the house trimming service that provided also be its output.Fig. 5 b shows the situation according to present embodiment, wherein noun node " House " is appointed as the input for house trimming service.Fig. 5 c illustration noun node " House " is appointed as the situation of the output of house trimming service, and Fig. 5 d illustration only noun node " House " is appointed as with verb node " Refurbish " and is associated, the notion/particular item that is contained by noun node " House " with indication is served with the house trimming that is provided in some unspecified mode to be associated.Attention: under the situation of house trimming, all above-mentioned links all are rational appointments.Based on this, the best option in the impossible present embodiment of the appointment of Fig. 5 a may be the illustrative appointment that is associated of Fig. 5 d.
In the present embodiment, data store 15 includes a plurality of records, and each record offers expection client's service corresponding to the association service provider.The contact details that is used to get in touch the provider and the description to the service that provided of writing with natural language are provided this record.
In addition, ontology server 10 also stores the corresponding lists of link to form index.In the present embodiment, this index utilizes certain suitable identifier to list to be stored in each record (so that can retrieve the record that is identified from data store 15) in the data store, has also listed together and linking that this record is associated.In the present embodiment, this link comprises noun node, verb node and relation (any in " ... input ", " ... output " or " being associated with ... ").In the present embodiment, preferably, require each provider to provide, form this information at the linking of each record by each record that just is associated with it.In order to help the provider to finish this task, provide for the read-only access that is stored in the body on the ontology server, and suitable navigation software, select to find only node to allow the provider to travel through this body.Alternatively, the ontology server keeper also can provide this information.This is for start-up system at first and to make its operation be useful.
Fig. 6 shows the part identical with the body shown in Fig. 4, and two frames 405,410, this two frames 405,410 expressions term that user by terminal 5 asks for instructions via the service search of the illustrative user interface input of Fig. 2 of controlling oneself.These two terms 405,410 are depicted as, and are respectively 1.0 by matching degree and are connected 421,422 with 0.48 coupling and match node 204 and 254.In addition, Fig. 6 comprises that in key component general (generic) " is matched with " junction symbol 420.In further detail the mode that the term that service search is asked for instructions is matched with the node in the body of storage is described hereinafter with reference to institute's drawings attached (in particular with reference to the process flow diagram of Fig. 7).
Therefore, referring now to Fig. 7, come in the method for one or more service log of paying close attention to of retrieval from data store 15 at the equipment that utilizes Fig. 1, first step is: the user utilizes illustrative user interface inputted search request in Fig. 2 at step S5.Then this service search request is sent to ontology server 10, ontology server 10 is further processed it according to following steps.
In step S10, this body is carried out at the term of the service search request that is received and is stored in name-matches between the title of node of the body in the ontology server.The purpose of this step is to make the user can freely key in this user-selected any term (these terms then are associated with the term that uses) in user interface in body, rather than directly select possible option from this body.Can use any appropriate method that is used to carry out this task.Yet,, in more detail the particular form that realizes this is in the present embodiment described below generally having described among Fig. 7 after the illustrative method.The result of name-matches step is definite: the verb node (A that all of body can partly be complementary with the verb of input service search request
i, 1≤i≤a wherein, a are the quantity (if present) of the verb node of coupling) and the verb of the verb node of each coupling mate the degree of correlation (CF (A
i)), and all noun node (P that can partly be complementary with the noun of input service search request of body
j, 1≤j≤p wherein, p are the quantity (if present) of the noun node of coupling) and the noun of the noun node of each coupling mate the degree of correlation (CF (P
j)).
After having finished step S10, method proceeds to step S15, determines that in this step S15 whether at least one noun node and at least one verb node all are complementary with the vocabulary of terms and the verb term of service search request respectively.If there be not the noun node or the verb node that can be complementary with the service search request, then this method proceeds to step S20, in this step S20, will beam back response to user terminal 5, to notify the user not find Search Results and to ask the user to use different search terms to attempt once more, this method finishes then.Yet, if mated at least one noun node (P
j) and at least one verb node (A
i), then this method proceeds to step S25.
In step S25, ontology server may make up by each that adopts coupling verb node and coupling noun node, and according to the relation (R) of the vocabulary of terms of in the original service searching request, expressing between the verb term these are linked at, form a plurality of services request (A through translation
i, P
j, R, CF (A
i), CF (P
j)).For example, if the service search request that the user imported is verb term " Get ", vocabulary of terms " Comm_Property ", and specified relationship is " ... input ", and have only single verb node (" Get " verb node 204) and single noun node (" CommercialProperty " node 254) to be complementary with it, then only formed single searching request, i.e. (A through translation
i=" Get ", P
j=" CommercialProperty ", R=" ... input ", CF (A
i)=1.0, CF (P
j)=0.48, i=a=j=p=1).Attention: the following calculating CF (A that when name-matches is discussed, illustrated in further detail
i) and CF (P
j) mode.
Generated in step S25 after the searching request of translation, this method proceeds to step S30, at this step S30 each in searching request and the index that is stored in the service chaining memory block 124 of translation each is linked to compare.Select the coupling mark to be confirmed as being higher than the link of predetermined threshold, this method proceeds to step S35 then.Below utilizing title is that the pseudo-code of " through the searching request and the link coupling of translation " has been set forth and carried out concrete mode relatively in the present invention.
After having finished step S30, this method proceeds to step S35, has determined whether to select in step S30 at least one link and associated record (actual recording storage is in data store 15) in this step S35.If be that then this method does not proceed to step S20, beams back response to user terminal 5 in this step S20, does not find Search Results to notify the user, and asks the user to use different search terms to attempt once more, this method finishes then.Yet, if in step S30, selected at least one link and associated record, then this method proceeds to step S40, in this step S40, from data store 15, retrieve the record of selected record or each selection, then with this as a result of the part of message send to user terminal 5, in user terminal 5 is in the result space 57 of the graphic user interface shown in Fig. 2 example, described result is shown to the user.
After step S40 finished, this method finished.
The details that to describe the name-matches step S10 in the present embodiment now and mate step S30 through the searching request and the link of translation.
Name-matches (step S10)
The purpose of this step be with the noun of searching request part and verb partly (freely be input to respectively in the text box 51 and 52 of user interface of Fig. 2 at user terminal) by the user be stored in ontology server 10 in body in corresponding noun node and verb node be complementary.In the present embodiment, this is to utilize three kinds of matched rules (direct matched rule, elementary cell name-matches rule and compound name matched rule) carry out, each matched rule adopts two term (A, B) as input, and output matching degree (CF (A, B)), if described term does not match according to this rule, this matching degree (CF (A then, B)) be 0, if described term matches to a certain degree, this matching degree (CF (A then, B)) be between 0 a value between (do not comprise significantly 0 itself) and 1 (comprise 1 itself), this value is that 1 expression is mated fully.
Directly matched rule compares two input terms simply, do not consider any punctuation mark, space etc., and the situation that letter occurs in these two terms (for example, upper case and lower case) difference in, output valve is 1 matching degree (CF (A if these terms are identical, B)), if different then output valve is 0 matching degree.
Elementary cell name-matches rule is not considered punctuate and capital and small letter etc. equally, and carries out in the following way: matching degree is initially set to 0, considers each following problem then successively:
1, first three of an input term letter identical (and order is identical)? if identical, then matching degree CF is added 0.3.
2, preceding four letters of input term identical (and order is identical)? if identical, then matching degree CF is added 0.3.
3, be first three of an input term letter identical (and order is identical) and the not last letter (being the not last letter of each input term) of input term also identical? if identical, then matching degree CF is added 0.3.
Thus, if in above-mentioned three problems neither one answer is yes, then matching degree will remain 0, and the result is not for matching.Answer is yes if only there is one in the problems referred to above (that is, only problem 1), and then will have matching degree is 0.3 coupling.If two (but not being whole three) in the described problem answer is yes (that is, and problem 1 and 2 or problem 1 and 3), then will have matching degree is 0.6 coupling.At last, if all three questions answer all are sure, then will have matching degree is 0.9 coupling.
When detecting two terms that will compare and all be compound name, used the compound name matched rule in the present embodiment.In this case, identify a plurality of constituent element elementary cell titles at each term, and attempt first constituent element elementary cell title of first term and the first constituent element elementary cell title of second term are mated, then second constituent element elementary cell title of first term and the second constituent element elementary cell title of second term are mated, or the like.Up to the end of the term that will have minimum constituent element constituent element elementary cell title with till corresponding constituent element elementary cell title in another term compares.Direct coupling is at first attempted in the trial of coupling constituent element elementary cell title, does not claim matched rule if find direct coupling then then use basic cell name.Coupling has been carried out in trial to constituent element elementary cell title after, calculate compound matching degree according to following formula:
CF (a wherein
i, b
i) be to use i among direct matched rule or determined compound term A of use elementary cell matched rule and the B to the matching degree of constituent element elementary cell title (if find coupling then equal 0); M is the quantity of the constituent element elementary cell title among the term A; N is the quantity of the constituent element elementary cell title among the term B; K is the smaller among m and the n.
After having set forth three types name-matches rule, can the algorithm that be adopted be described with following pseudo-code:
Comment：first?process?the?verb?term，A，of?the?service?search?request；
FOR?each?verb?node，B＝b
1，b
2，...，b
k，in?the?stored?ontology{
TRY?to?find?direct?match
IF?successful?record?match；
NEXT?verb?node；
END?IF
IF?A?and?b
index?are?atomic?names
TRY?atomic?name?matching
IF?successful?record?match；
NEXT?verb?node；
END?IF
ELSE?IF?A?and?b
index?are?both?compound?names
TRY?compound?name?matching
IF?successful?record?match；
NEXT?verb?node；
END?IF
END?IF
END?FOR
Comment：Then?repeat?forthe?noun?term，O，of?the?service?search?request
FOR?each?noun?node，P＝p
1，p
2，...，p
i，in?the?stored?ontology
TRY?to?find?direct?match
IF?successful?record?match；
NEXT?verb?node；
END?IF
IF?O?and?p
index?are?atomic?names
TRY?atomic?name?matching
IF?successful?record?match；
NEXT?verb?node；
END?IF
ELSE?IF?O?and?p
index?are?both?compound?names
TRY?compound?name?matching
IF?successful?record?match；
NEXT?verb?node；
END?IF
END?IF
END?FOR
Above pseudo-chip segment is represented in fact: at first obtain the verb term of user's input, then all verb nodes that are stored in the body are circulated to search coupling.In the each iteration of round-robin, at first search direct coupling, if the coupling of finding, then by would writing down the fact of this coupling in the local storage list that clauses and subclauses is placed into the noun node that comprises coupling and matching degree.If directly do not mate, then judge the verb term imported and current verb node whether all be the elementary cell title (in the present invention, detect compound name in the following manner: search in punctuation mark space, underscore, hyphen, fullstop, oblique line, colon, comma or the branch that separates two alphabetic strings, or the capital and small letter of (not comprising initial) (for example changes in the middle of the alphabetic string, Comm_property, CommProperty, Comm Property)); If the elementary cell title then uses elementary cell name-matches rule to search coupling, and if find coupling, the fact of this coupling then write down as described above.If verb term of being imported and current verb node all are compound name (as mentioned above), then these titles are separated into constituent element elementary cell title separately, and use the compound name matched rule to search coupling, if the coupling of finding then writes down the fact of this coupling as described above.If when this process finishes, do not find coupling, current iteration is finished, and do not write down any coupling, and utilize next verb node to start new iteration.
Attention: in above-mentioned algorithm, if the verb term of being imported is the elementary cell title, and current verb node is compound, and perhaps situation is opposite, then can not find coupling (except being under the exception that finds direct coupling).It is generally acknowledged that this is not a problem, this is owing to attempt mating compound name and another compound name etc. normally better.However, unless the situation that two titles all are the elementary cell titles etc., alternative embodiment can be operated by the mode of alternative by using the compound name matched rule all the time.
Under necessary situation about changing in addition, be used for that the vocabulary of terms of service search request carried out Matching Algorithm with the noun node of body and be used for the verb term and the verb algorithm of nodes is identical.
Searching request and link coupling through translation
On the whole, carry out this process in the present embodiment as follows.Each service search request through translation is considered successively.Under the situation of using body,, form the subtree of the action node of searching request through translating by comprising as the similar or subclass (comprising subclass etc.) of action node and all nodes of action node itself.Whether then each clauses and subclauses that are stored in the concordance list in the ontology server 10 are checked, be a node in this subtree to judge its action node.If, evaluate matches degree in the following manner then: considered vocabulary of terms and respectively in the relation in the link information of the searching request of translation and storage.Then a matching degree and a threshold value of being assessed compared,, then select corresponding service log at retrieval with to sending of user terminal if matching degree exceeds this threshold value.
Below set forth the concrete mode of the matching degree between the linking in the subtree of action node that searching request and its action node through translation fall into the searching request through translating.Generally, at first check to judge whether noun node in this link falls in the subtree through the noun node of the searching request of translation (under necessary situation about changing, to derive with same way as at the action subtree) in addition.If do not fall into, then matching degree is made as 0, and coupling finishes.Otherwise, will compare with the relation that this links through the relation of searching request of translation, if the two is identical, will concern that then comparison score is made as 1.If one relation is " being associated with ... " and another is not (that is, another is " ... input " or " ... output "), will concern that then comparison score is made as 0.5.If relation is " ... input " and another is " ... output ", will concern that then comparison score is made as 0.At last, use following formula to calculate the value of matching degree:
Matching degree=(CF (A)+CF (P)+CF (R))/(6-(CF (A)+CF (P)+CF (R)))
Wherein, CF (A) is the matching degree between the verb node of verb term and the current searching request of examining or check through translating of input, CF (P) is the matching degree between the noun node of vocabulary of terms and the current searching request of examining or check through translating of input, and CF (R) is the aforesaid comparison score (its value is 0,0.5 or 1) that concerns.Therefore, note: if CF (A)=CF (P)=CF (R)=1, then matching degree=1; If CF (A)=CF (P)=CF (R)=0, then matching degree=0; And if CF (A)=CF (P)=CF (R)=0.5, then matching degree=1/3.
In the present embodiment, threshold value is made as 0.4.Yet, in alternative embodiment, can select any matching degree assessment greater than 0, in fact finally selected the individual record of choosing of limited quantity (for example 10), and they have been sent to user terminal.In either case, record is chosen in the order demonstration that preferably reduces by matching degree assessment mark.
The utilization example
For the above-mentioned discussion of illustration, will examine or check the request of example inputted search now.For this illustrative purpose, suppose and in data store 15, only have two service logs, wherein has following association link (by provider's input): for first record, be verb node " Sell " noun node " House " and relation " ... output " (provider in this case is the real estate agents merchant that the service of selling house is provided to expection house buyer); And for second record, be verb node " Buy ", noun node " Motel " and relation " ... input " (provider in this case is big company of motel, and it is interested that it buys motel to the owner of motel from the motel of attempting sell them).
User's inputted search request " Get " (being input in the verb text box 52) and " Comm_Property " (being input in the noun text box 51), and select " ... input " radio button 53.Resulting searching request is sent to ontology server, in the name-matches of the execution in step S10 of this ontology server place.In this step, searching request term " get " and verb node " Get " directly mate and do not have other couplings.Searching request term " Comm_Property " does not directly mate with any noun node.(by the existence of underscore character) determines it is compound name, and it is matched noun node " Commercial Property " with matching degree CF (P)=(0.6+1)/(2+2-(0.6+1))=2/3 ≈ 0.67 (seeing above-mentioned compound matching formula) by the compound name matched rule, in view of the answer of problem 1 and problem 2 in the elementary cell name-matches rule all is sure, the first elementary cell title " Comm " is mated together with " Commercial ", and matching degree is 0.6.Yet, not together with any other noun node matching in searching request term " Comm_Property " and the body.
This method proceeds to step S25 then, wherein utilizes verb node " Get ", noun node " Commercial Property ", relation " ... input ", CF (A)=1 and CF (P)=0.67 to generate single searching request through translation.
This method proceeds to step S30 then, wherein attempts will mating with a record that is stored in the data store 15 through the searching request of translation with the foundation chained list.Because verb node " Sell " is not in the subtree of verb node " Get ", so at first record (" Sell ", " House " ... output) link be unmatched.Yet because verb node " Buy " is in the subtree of " Get ", and noun node " Motel " is in the subtree of " Commercial Property ", so at second record (" Buy ", " Motel " ... input) link mate.
Matching degree=(1+0.67+1)/(6-(1+0.67+1))=0.80
Since in the present embodiment threshold value is made as 0.4, thus this record selected at step S40, and send it back to user terminal 5.
Therefore, in short, particularly with reference to Fig. 6, present embodiment provides a kind of service description document has been stored in method in the computerized storage system, wherein each document is associated with at least one this body node of verb 204 and at least one this body node of noun 254, each this body node of verb has with one or more of other these body nodes of verb and links, and each this body node of noun has with one or more of other these body nodes of noun and links, thereby these verb nodes form verb space 200, and these noun nodes form noun space 250 (this verb space and this noun space are unique limited bodies), and present embodiment also provide a kind of from a plurality of service description document of storage in this way retrieval service the method for document is described, this method may further comprise the steps:
The control user interface, asking at least one verb request term 405 and at least one noun request term 410 from the user,
Each verb request term 405 is associated with corresponding verb node 204, and each noun request term 410 is associated with corresponding noun node 254,
The verb node 204 that each is corresponding compares with each verb node 212,214 that is associated with each service description document of storing,
The noun node 254 that each is corresponding compares with each noun node 262,266 that is associated with each service description document of storing,
Select retrieving zero or more a plurality of service description document of being stored based on described comparison step, and the control user interface makes described user can retrieve one or more selected document to notify user-selected document.
Modified example
Replacement with chained list and at the index stores of stored record on ontology server, can also with this information stores in different positions, for example be stored in the same data store identical with stored record itself.In fact, link information can only be the part of data recording itself, mates the time that is spent with being associated with to carry out between the linking of the data recording of being stored although this is increased in possibly through the searching request of translation.
Replace carrying out in the above described manner the elementary cell name-matches, in alternative embodiment, can also test a large amount of different rules to have and aforementioned different coupling mark,, replacement do not consider success or failure but testing at each rule, can also carry out test in the following manner: from having the test of highest score, and having the end of test (EOT) of lowest fractional, and, just stop to carry out other test as long as a test is sure.For example, can test three following rules:
1, is first three of an input term letter identical (and order is identical) and the last letter (being the not last letter of each term) of input term also identical? if identical, then matching degree CF is made as 0.5, and finishes the elementary cell coupling, otherwise
2, preceding four letters of input term identical (and order is identical)? if identical, then matching degree CF is made as 0.5, and finishes the elementary cell coupling, otherwise
3, preceding 3 letters of input term identical (and order is identical)? if identical, then matching degree CF be made as 0.3, and finish the elementary cell coupling.
In the above-described embodiments, the compound name matching algorithm is operated in the following manner: first elementary cell title of first compound word and the first elementary cell title of second compound word are compared, second elementary cell title of first compound word and the second elementary cell title of second compound word are compared, by that analogy.If the ordering between compound name is different, even for the compound name of sharing a large amount of elementary cell titles, this also may can not find coupling.The compound name matched rule of the alternative that can be used for addressing this problem has below been described:
If A={A1 ..., Am}, B={B1 ..., Bn} is two compound names, wherein respectively, and A1 ... Am is a m elementary cell title at A, and B1 ..., Bn is a n elementary cell title at B.If C={C1 ..., Ck} is k the elementary cell title of between A and B mating, while CF (C)=CF (C1) ..., CF (Ck) } and be matching degree.Can be by each the elementary cell title in first term and each the elementary cell title in second term be mated, and judge that based on this result which elementary cell title paired with each other is used to form total compound name matching result, calculates matching degree CF (C).In form, we have following algorithm:
FOR?each?term?Ai∈A＝{A1，…Am}
CF(Ai，B)＝0
bmax＝0
FOR?each?term?Bj∈B＝{B1，…，Bn}
IF CF(Ai，Bj)＞CF(Ai，B) THEN
CF(Ai，B)＝CF(Ai，Bj)
bmax＝j
ENDIF
ENDFOR
IF?CF(Ai，B)＞0 THEN
C＜-C+(Ai，Bbmax)
A＜-A-Ai
B＜-B-Bbmax
ENDIF
ENDFOR
This algorithm is operated in the following manner: each elementary cell title of testing the first term A at each elementary cell title of the second term B; Pairing with maximum match degree is added among the C, and respectively from A and B with its deletion.Till the time that this algorithm finishes, C includes all pairings from the coupling of A and B.
Thus with the matching degree CF between A and the B (A B) is calculated as follows:
Claims (13)
1, a kind of method that the document of a plurality of storages is searched for, this method may further comprise the steps:
Store described a plurality of document;
The expression of storage body, described body comprise a plurality of nodes and be divided at least two unique subspaces of being mutually related;
At in described a plurality of documents each, at least one of the node of storage and first unique subspace of body is related, with related with at least one of the node of second unique subspace of body;
The control user interface so that the user can use free text input to import nearly at least two search terms, and can be searched for each term and is associated with corresponding unique subspace of body;
The search term of each input is only compared with the node of corresponding subspace, to attempt determining that one or more may mate or the part coupling; And
May mate or the node of part coupling based on each, and related between the node of the document of described storage and body, select the document of one or more described storage, to present to the user.
2, a kind of method of storing a plurality of electronic documents may further comprise the steps:
Generate related with at least one of the node of first kind node at each electronic document, with related with at least one of the node of second type node, described node belongs to predetermined body, and this predetermined body has following characteristic: the subtree of the node of given type only comprises the node of same given type;
With, according to can making described association easily be linked to the mode of corresponding document, the association that generates with particular document and at described document to or associated group be stored in the number storage.
3, method according to claim 1 and 2, wherein, first subspace comprises the verb node or first kind node is the verb node, and second subspace comprises the noun node or second type node is the noun node.
4, according to claim 1,2 or 3 described methods, wherein, with described association be associated with each related to or the sign of the described document of associated group be stored in the index that is used for searching for efficiently together.
5, according to the described method of arbitrary aforementioned claim, wherein, described document comprises the natural language description to service.
6, according to the described method of arbitrary aforementioned claim, further comprising the steps of: generate relation identifier, it identifies in may concerning one of a limited quantity independence between the node in the node or the first kind and second subspace or second type first the subspace in; With with described relation identifier with described related to or associated group store together.
7, a kind of search method, be used for from storing one or more electronic document of electronic storage device retrieval of a plurality of electronic documents, described document is stored according to claim 2 or the method that is subordinated to arbitrary claim of claim 2, and described search method may further comprise the steps:
Receive the electronic signal of expression searching request, described searching request comprises at least one first term that is associated with the node of the first kind of being scheduled to body and at least one second term that is associated with the node of second type of predetermined body;
The node of first term and a plurality of described first kind is compared, and the node of second term and a plurality of described second types is compared, and under the situation of determining at least a portion coupling, distribute matching degree for each this node;
Generate at least one searching request through translation, described searching request comprises the described matched node of at least one described first kind, the described matched node of at least one described second type, and the matching degree that is associated with each node;
With each each matched node of searching request through translation, with by corresponding to the storage of the electronic document of each storage related to or the corresponding node of the same type that associated group identified compare;
Based on described searching request and described storage through translation related to or associated group between comparative result, select to retrieve the document of usefulness; And
The electronic signal of each electronic document of choosing is represented or is identified in output.
8, a kind of searching request generation method is used for being created on the searching request of using according to the method for claim 7, and described searching request generation method may further comprise the steps:
The control user interface is to ask first term from the user;
The control user interface is to ask second term from the user;
The control user interface is selected a plurality of in may relationship types one with the request user, expresses the relation between first term and second term; And
Information based on user's input generates searching request.
9, a kind ofly be used to store and the equipment of retrieve electronic document, comprise:
Electronic data storage portion comprises the device that is used to store a plurality of electronic documents;
Another electronic data storage device structure, be used to store node with at least one second type of the node of each electronic document and at least one first kind of predetermined body and predetermined body is associated related to or associated group;
Request generating device is used to generate the searching request that comprises first term and second term;
Translating equipment, be used for generating in the following manner searching request through translation, promptly, first term of searching request and the node of the first kind are compared, and second term of searching request and the node of second type are compared, find and the corresponding specific node of the term of searching request; And
Comparison means, be used for each through searching request and each storage of translation related to or associated group compare, and select it has been determined fully near the document that mates.
10, equipment according to claim 9, wherein, electronic data storage portion also comprises described another electronic data storage device structure.
11, the electronic data storage portion that uses of a kind of equipment that is used in claim 8 or 9, the a plurality of electronic documents of described data storage stores, with node with the node of each electronic document and at least one first kind of predetermined body and at least one second type be associated related to or associated group.
12, a kind of enforcement of rights that is used for requires 1 to 7 each the computer program of method.
13, the carrier arrangement of the computer program of carrying claim 12.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| GB0320205.8 | 2003-08-28 | ||
| GBGB0320205.8A GB0320205D0 (en) | 2003-08-28 | 2003-08-28 | Method and apparatus for storing and retrieving data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN1846210A true CN1846210A (en) | 2006-10-11 |
Family
ID=28686513
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNA2004800248868A Pending CN1846210A (en) | 2003-08-28 | 2004-08-20 | Method and apparatus for storing and retrieving data using ontologies |
Country Status (6)
| Country | Link |
|---|---|
| US (1) | US20060248458A1 (en) |
| EP (1) | EP1661045A1 (en) |
| CN (1) | CN1846210A (en) |
| CA (1) | CA2536760A1 (en) |
| GB (1) | GB0320205D0 (en) |
| WO (1) | WO2005022408A1 (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103548023A (en) * | 2011-05-27 | 2014-01-29 | 国际商业机器公司 | Automated self-service user support based on ontology |
| CN101996208B (en) * | 2009-08-31 | 2014-04-02 | 国际商业机器公司 | Method and system for database semantic query answering |
| CN110297883A (en) * | 2018-03-22 | 2019-10-01 | 卡西欧计算机株式会社 | Information display device, information display method and recording medium |
Families Citing this family (30)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7373355B2 (en) * | 2004-09-03 | 2008-05-13 | Metallect Corporation | System and method for relating applications in a computing system |
| US8677377B2 (en) * | 2005-09-08 | 2014-03-18 | Apple Inc. | Method and apparatus for building an intelligent automated assistant |
| WO2007084791A2 (en) | 2006-01-20 | 2007-07-26 | Glenbrook Associates, Inc. | System and method for managing context-rich database |
| US8495004B2 (en) | 2006-03-27 | 2013-07-23 | International Business Machines Corporation | Determining and storing at least one results set in a global ontology database for future use by an entity that subscribes to the global ontology database |
| US8204891B2 (en) * | 2007-09-21 | 2012-06-19 | Limelight Networks, Inc. | Method and subsystem for searching media content within a content-search-service system |
| US8396878B2 (en) | 2006-09-22 | 2013-03-12 | Limelight Networks, Inc. | Methods and systems for generating automated tags for video files |
| US8966389B2 (en) * | 2006-09-22 | 2015-02-24 | Limelight Networks, Inc. | Visual interface for identifying positions of interest within a sequentially ordered information encoding |
| US7917492B2 (en) * | 2007-09-21 | 2011-03-29 | Limelight Networks, Inc. | Method and subsystem for information acquisition and aggregation to facilitate ontology and language-model generation within a content-search-service system |
| US9015172B2 (en) | 2006-09-22 | 2015-04-21 | Limelight Networks, Inc. | Method and subsystem for searching media content within a content-search service system |
| US7886334B1 (en) | 2006-12-11 | 2011-02-08 | Qurio Holdings, Inc. | System and method for social network trust assessment |
| US8346864B1 (en) | 2006-12-13 | 2013-01-01 | Qurio Holdings, Inc. | Systems and methods for social network based conferencing |
| US8135800B1 (en) | 2006-12-27 | 2012-03-13 | Qurio Holdings, Inc. | System and method for user classification based on social network aware content analysis |
| US9195996B1 (en) * | 2006-12-27 | 2015-11-24 | Qurio Holdings, Inc. | System and method for classification of communication sessions in a social network |
| US8140535B2 (en) * | 2007-10-23 | 2012-03-20 | International Business Machines Corporation | Ontology-based network search engine |
| US8041702B2 (en) * | 2007-10-25 | 2011-10-18 | International Business Machines Corporation | Ontology-based network search engine |
| US8135730B2 (en) * | 2009-06-09 | 2012-03-13 | International Business Machines Corporation | Ontology-based searching in database systems |
| US8498986B1 (en) * | 2012-01-31 | 2013-07-30 | Business Objects Software Ltd. | Classifying data using machine learning |
| USD754161S1 (en) | 2012-11-26 | 2016-04-19 | Nero Ag | Device with a display screen with graphical user interface |
| US20140149932A1 (en) * | 2012-11-26 | 2014-05-29 | Nero Ag | System and method for providing a tapestry presentation |
| US10140638B2 (en) * | 2012-12-06 | 2018-11-27 | International Business Machines Corporation | Providing information technology resiliency in a cloud-based services marketplace |
| CN103631948B (en) * | 2013-12-11 | 2017-01-11 | 北京京东尚科信息技术有限公司 | Identifying method of named entities |
| CN103838883A (en) * | 2014-03-31 | 2014-06-04 | 上海久科信息技术有限公司 | Intelligent SKU matching method |
| US20160162458A1 (en) * | 2014-12-09 | 2016-06-09 | Idibon, Inc. | Graphical systems and methods for human-in-the-loop machine intelligence |
| US11157260B2 (en) | 2015-09-18 | 2021-10-26 | ReactiveCore LLC | Efficient information storage and retrieval using subgraphs |
| US9372684B1 (en) * | 2015-09-18 | 2016-06-21 | ReactiveCore LLC | System and method for providing supplemental functionalities to a computer program via an ontology instance |
| US9864598B2 (en) | 2015-09-18 | 2018-01-09 | ReactiveCore LLC | System and method for providing supplemental functionalities to a computer program |
| US10599718B2 (en) * | 2015-10-09 | 2020-03-24 | Software Ag | Systems and/or methods for graph based declarative mapping |
| US20170344896A1 (en) * | 2016-05-31 | 2017-11-30 | Deere & Company | Disambiguation of agricultural input |
| US11537558B2 (en) * | 2018-11-13 | 2022-12-27 | Dokkio, Inc. | File management systems and methods |
| US11379430B2 (en) | 2018-11-13 | 2022-07-05 | Dokkio, Inc. | File management systems and methods |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| AU6849196A (en) * | 1995-08-16 | 1997-03-19 | Syracuse University | Multilingual document retrieval system and method using semantic vector matching |
| US6076088A (en) * | 1996-02-09 | 2000-06-13 | Paik; Woojin | Information extraction system and method using concept relation concept (CRC) triples |
| WO2000005664A1 (en) * | 1998-07-24 | 2000-02-03 | Jarg Corporation | Search system and method based on multiple ontologies |
| US20020078090A1 (en) * | 2000-06-30 | 2002-06-20 | Hwang Chung Hee | Ontological concept-based, user-centric text summarization |
| WO2002005137A2 (en) * | 2000-07-07 | 2002-01-17 | Criticalpoint Software Corporation | Methods and system for generating and searching ontology databases |
| US6675159B1 (en) * | 2000-07-27 | 2004-01-06 | Science Applic Int Corp | Concept-based search and retrieval system |
| US6766316B2 (en) * | 2001-01-18 | 2004-07-20 | Science Applications International Corporation | Method and system of ranking and clustering for document indexing and retrieval |
-
2003
- 2003-08-28 GB GBGB0320205.8A patent/GB0320205D0/en not_active Ceased
-
2004
- 2004-08-20 WO PCT/GB2004/003583 patent/WO2005022408A1/en not_active Application Discontinuation
- 2004-08-20 CA CA002536760A patent/CA2536760A1/en not_active Abandoned
- 2004-08-20 US US10/568,825 patent/US20060248458A1/en not_active Abandoned
- 2004-08-20 EP EP04768140A patent/EP1661045A1/en not_active Ceased
- 2004-08-20 CN CNA2004800248868A patent/CN1846210A/en active Pending
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101996208B (en) * | 2009-08-31 | 2014-04-02 | 国际商业机器公司 | Method and system for database semantic query answering |
| CN103548023A (en) * | 2011-05-27 | 2014-01-29 | 国际商业机器公司 | Automated self-service user support based on ontology |
| US10019512B2 (en) | 2011-05-27 | 2018-07-10 | International Business Machines Corporation | Automated self-service user support based on ontology analysis |
| US10037377B2 (en) | 2011-05-27 | 2018-07-31 | International Business Machines Corporation | Automated self-service user support based on ontology analysis |
| US10162885B2 (en) | 2011-05-27 | 2018-12-25 | International Business Machines Corporation | Automated self-service user support based on ontology analysis |
| CN110297883A (en) * | 2018-03-22 | 2019-10-01 | 卡西欧计算机株式会社 | Information display device, information display method and recording medium |
| CN110297883B (en) * | 2018-03-22 | 2023-10-31 | 卡西欧计算机株式会社 | Information display device, information display method, and recording medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CA2536760A1 (en) | 2005-03-10 |
| GB0320205D0 (en) | 2003-10-01 |
| WO2005022408A1 (en) | 2005-03-10 |
| US20060248458A1 (en) | 2006-11-02 |
| EP1661045A1 (en) | 2006-05-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1846210A (en) | Method and apparatus for storing and retrieving data using ontologies | |
| Feng et al. | The deep learning–based recommender system “Pubmender” for choosing a biomedical publication venue: Development and validation study | |
| US8335787B2 (en) | Topic word generation method and system | |
| US7984035B2 (en) | Context-based document search | |
| CN1284107C (en) | Information storage and retrieval | |
| CN1694100A (en) | Content propagation for enhanced document retrieval | |
| CA2789010C (en) | Propagating classification decisions | |
| CN1489089A (en) | Document search system and question answer system | |
| US20110191310A1 (en) | Method and system for ranking intellectual property documents using claim analysis | |
| CN1871603A (en) | System and method for processing a query | |
| CN101044481A (en) | A method, system, and computer program product for searching for, navigating among, and ranking of documents in a personal web | |
| Sang et al. | Learn to personalized image search from the photo sharing websites | |
| CN101055580A (en) | System, method and user interface for retrieving documents | |
| US20120130999A1 (en) | Method and Apparatus for Searching Electronic Documents | |
| Remi et al. | Domain ontology driven fuzzy semantic information retrieval | |
| Chan et al. | Automated online news classification with personalization | |
| Sharifpour et al. | Large-scale analysis of query logs to profile users for dataset search | |
| Bashir et al. | Opinion-Based Entity Ranking using learning to rank | |
| Cecchini et al. | Multiobjective evolutionary algorithms for context‐based search | |
| Wu et al. | Document keyphrases as subject metadata: incorporating document key concepts in search results | |
| US20140095465A1 (en) | Method and apparatus for determining rank of web pages based upon past content portion selections | |
| Williams et al. | Classifying and ranking search engine results as potential sources of plagiarism | |
| US20140149378A1 (en) | Method and apparatus for determining rank of web pages based upon past content portion selections | |
| Naveed et al. | Feature sentiment diversification of user generated reviews: the FREuD approach | |
| Langnickel et al. | Pre2Pub—tracking the path from preprint to journal article: algorithm development and validation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |